sign in sign up
<<
Home , Desmodromic valve, Engine configuration, Harry Ricardo, Junk head, Napier Nomad, Napier Sabre

Citation for published version: Turner, J, Head, RA, Chang, J, Engineer, N, Wijetunge, RS, Blundell, DW & Burke, P 2019, '2- Engine Options for Automotive Use: A Fundamental Comparison of Different Potential Scavenging Arrangements for Medium-Duty Truck Applications', SAE Technical Paper Series, pp. 1-21. https://doi.org/10.4271/2019-01-0071

DOI: 10.4271/2019-01-0071

Publication date: 2019

Document Version Peer reviewed version

Link to publication

The final publication is available at SAE Mobilus via https://doi.org/10.4271/2019-01-0071

University of Bath

Alternative formats If you require this document in an alternative format, please contact: [email protected]

General rights Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights.

Take down policy If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.

Download date: 27. Sep. 2021 Paper Offer 19FFL-0023 2-Stroke Engine Options for Automotive Use: A Fundamental Comparison of Different Potential Scavenging Arrangements for Medium-Duty Truck Applications

Author, co-author (Do NOT enter this information. It will be pulled from participant tab in MyTechZone) Affiliation (Do NOT enter this information. It will be pulled from participant tab in MyTechZone)

Abstract For the opposed- engine, once the port timing obtained by the optimizer had been established, a supplementary study was conducted looking at the effect of relative phasing of the The work presented here seeks to compare different means of on performance and economy. This was found to have a small effect providing scavenging systems for an automotive 2-stroke engine. It on fuel consumption for a significant change in , follows on from previous work solely investigating uniflow suggesting that, if available, variable phasing could be a scavenging systems, and aims to provide context for the results very important control actuator for compression ignition in discovered there as well as to assess the benefits of a new scavenging such an engine. system: the reverse-uniflow sleeve-. Importantly, it was found that existing experiential guidelines for port For the study the general performance of the engine was taken to be angle-area specification for loop-scavenged, piston-ported engines suitable to power a medium-duty truck, and all of the concepts using compression could also be applied to all of the other discussed here were compared in terms of indicated fuel consumption scavenging types, this having been done here in order to provide a for the same swept volume using a one-dimensional engine starting point for the work. This important result has not been simulation package. In order to investigate the sleeve-valve designs demonstrated before for such a wide range of architectures. The layout drawings and analysis of the Rolls-Royce Crecy-type sleeve optimizer employed then allowed further improvements to be made had to be undertaken. over the starting point. The paper therefore presents a fundamental comparison of scavenging systems using a new approach, providing A new methodology for optimization was developed and the analysis insights and information which have not been shown before. process also took into account work done by the charging system, this being assumed to be a combination of and to permit some exhaust waste heat recovery. Introduction

As a result of this work it was found that the opposed-piston Context and the need for greater investment in configuration provides the best attributes since it allows maximum combustion engine technology expansion and minimum heat transfer. It gave net specific fuel consumption results which were 9.6% lower than the loop-scavenged Automotive transportation has helped to revolutionize society and engine (which was marginally the worst of the configurations thus in many ways the internal combustion engine (ICE) helped to investigated). The other uniflow systems were next, with the reverse define the twentieth century. However, the synergistic development being the most promising (3.4% better than the loop- of the ICE with fossil fuels has resulted in the emission of large scavenged engine). quantities of carbon dioxide which, because it is a greenhouse gas, contributes to global warming. It is imperative that all measures are Furthermore, although the general performance the loop-scavenged taken to reduce the fossil CO2 impact of transportation while still configuration was closer to the other designs than was initially providing the economic benefits of affordable transportation that the expected, it was found to be compromised by its requirement to have ICE has brought. While many commentators consider that future and exhaust ports at the same height in the cylinder, thus ground transportation should consequently be provided solely by lengthening the gas exchange events for any given angle-area and battery electric vehicles and the hydrogen proton exchange consequently reducing the effective (or trapped) compression and membrane (PEM) (FC), due to the opportunity that these expansion ratios. This was despite the use of a charge trapping valve provide to decarbonize their energy supply through renewable to provide asymmetric port timing and minimize charge short- electricity generation, there is considerable belief within industry and circuiting, the adoption of which was felt to be a factor in its better- academia that due to the cost involved – both in terms of than-expected performance. Finally, the reverse-loop-scavenged infrastructure and to the vehicle purchaser – this would take many poppet-valve type was found to be so compromised by breathing and years to complete. valve train kinematics that it was not taken to a full optimization. Conversely, there is unarguably considerable potential left in the internal combustion engine (ICE) with regards to improving its Page 1 of 21

7/20/2015 efficiency. This statement is made simply because whereas the production volumes) of engines designed and developed for all efficiency of an electric motor is in the region of 93-97% across the vehicular applications, one might observe that in fact the 4-stroke majority of its operating map, the peak thermal efficiency of a typical engine is in fact an automotive peculiarity. passenger or medium truck combustion engine is only in the region of 37-45%. This means that logically there is a greater Sher, in reviewing 2-stroke scavenging, notes that all engines prior to potential to reduce losses in the ICE, even though it is limited by the Nikolaus Otto’s 4-stroke device were 2-strokes [1]. Sir Dugald Clerk efficiency of the Carnot cycle whereas the electric motor and fuel cell patented what may be considered the first commercially successful 2- (FC) are not. Even though the energy storage system for an ICE stroke engine in 1881. This was what would now be termed a boasts an efficiency of 100% (disregarding its minimal evaporative reverse-uniflow engine, utilizing a one-way inlet valve in the head emissions) – versus about 85% for a battery – in terms of energetic and piston-controlled exhaust ports [1]. Sir ’s later efficiency the ICE system cannot get back on par with that of an Dolphin engine was similar to this, albeit with a slightly different (EV). On the other hand, the amount of energy (and valving mechanism. It was Joseph Day who, together with one of his with it the achievable range of the vehicle) that can easily be stored in workmen, Frederick Cook, developed the piston-ported 2-stroke hydrocarbon form on board a vehicle dwarfs that which a battery can engine in Bath in 1889-1891 [2]. Allegedly, and like the designs of currently hold, and this state of affairs will continue until some Clerk, this was to circumvent the Otto 4-stroke engine patents of significant breakthroughs in battery chemistry are made. These, if 1876. Notwithstanding this, the 2-stroke cycle was not embraced they are ever realized, will then take of the order of two decades to with enthusiasm by the automotive industry; one can imagine that bring to mass production. this was because the Otto cycle was much simpler to comprehend and optimize within the understanding of engines at that time because of A further issue is that whereas the ICE has historically been proven to its meaningful separation of cycle events. This situation is true be a silver bullet for all forms of transportation, alternative whether the engine employs spark-ignition (SI) or compression- propulsion systems do not have the same potential reach. Indeed, ignition (CI) combustion. heavier means of surface transportation such as long-distance haulage and shipping may never be able to adopt these solutions (mainly due As the engineering science pertaining to the thermodynamics of to the challenges of energy storage). This is even more likely for combustion engines has developed it has become apparent that the 2- long-distance commercial aviation, despite efforts being made to stroke cycle does in fact possess some very significant benefits, produce battery-electric and PEM FC-powered light aircraft. especially versus the simple throttled SI 4-stroke engine. These advantages can be summarized as: Against this backdrop it is obvious that technologies that improve the efficiency of all forms of the ICE ought to be seen to be of crucial 1. The minimization of pumping work, through the elimination of importance for the foreseeable future. This is true regardless of the a dedicated induction stroke. Instead of the induction process type or degree of electrical hybrization that may be applied to it. being undertaken at the same expansion ratio as the combustion When it is further considered that the vehicle original equipment part of the cycle, the 2-stroke is free to adopt scavenge manufacturers (OEMs) will undoubtedly require a portfolio approach arrangements more-optimized for this purpose and so mitigate to comply with emissions legislation during any interim period of this loss. migration to a fully-electric future, improvements to the ICE will also ensure their continued relevance for an ever-longer time period. Due 2. A potential for reduced friction. This is mainly due to two to their incomparable cost-effectiveness for all stakeholders in the things: the potential for a simpler mechanism for the gas transportation economic system, this in turn strengthens the business exchange process (through the removal of a requirement for a case for investing in ICE technology. half-engine-speed drive for this purpose) and increased mechanical efficiency, because The 2-stroke engine in transportation 퐵푀퐸푃  = Eqn 1 In the automotive world the reciprocating 4-stroke engine has utterly 푀푒푐ℎ푎푛𝑖푐푎푙 퐼푀퐸푃 dominated all other forms of prime mover to date. The reality is that it was the invention of this type of engine by Nikolaus Otto, together Where 푀푒푐ℎ푎푛𝑖푐푎푙 is the mechanical efficiency, BMEP is with its utility, which gave rise to the automobile (and not the other brake and IMEP is the indicated mean way round). In comparison, except for the essentially separate effective pressure. markets of cost-effective and high-performance motorcycles, the 2- stroke engine has historically been neglected for light- and medium- 3. The potential for simpler and lighter construction. This is again duty applications (although it must be remembered that numerous related to the observations on gas exchange mechanisms, and cheap were built using very simple 2-stroke engines in Eastern has ramifications in terms of cost and engine mass. It is also Europe until the 1980s). In the West even the was related to the specific power increase possible and the fact that more successful than the reciprocating 2-stroke engine in passenger this can be traded off against the BMEP necessary, resulting in cars, with Mazda producing significant (if still small in comparison to a requirement for lower peak cylinder pressures and hence their 4-stroke engine volumes) numbers of Wankel-type rotary lighter scantlings of the engine. engines until production of that line of vehicles ceased with the RX-8 in 2012. 4. Related to 3. above, for a given swept volume, at any operating condition the BMEP required of the 2-stroke engine is half that This is the case in the automotive world. However, in areas where its 4-stroke counterpart. This reduced in-cylinder load leads to either power density or efficiency were the primary goals for a reduced peak cycle temperatures and hence lower emissions of particular engine application the 2-stroke cycle reigned, and oxides of nitrogen (NOx). For similar reasons the thermal continues to reign, supreme: the largest and smallest reciprocating losses are lower. engines operate on this cycle. Considering the specific types (not the Page 2 of 21

7/20/2015 For these reasons, the 2-stroke engine has been very successful in from the positive piston motion in the pumping strokes there are applications where maximum efficiency has been of overriding challenges with fuel-air mixing. The large quantity of residual gas importance or where its simplicity and power density advantages can that is trapped, however, does mean that the 2-stroke is a natural be realized within the prevailing emissions legislation (for example, arena in which to apply HCCI-type combustion systems, and this has in , where NOx limits are relatively high, and historically been successfully demonstrated in a variety of such portable equipment, where cost is an over-riding imperative, engines by Honda and others [13,14,15]. The use of these types of respectively). Although it has never achieved significant impact in combustion system, together with direct injection (DI), therefore the automotive market, it has periodically been investigated as an provide a feasible route to low pollutant emissions and high alternative to the 4-stroke engine for on-road applications; for efficiency in the 2-stroke engine. This work seeks to investigate the example, many OEMs researched piston-ported 2-strokes in the efficiency levels that can be achieved when different scavenging 1980s as a result of work undertaken by Orbital Engine Corporation mechanisms are used for the 2-stroke engine when it operates on just [3,4] and Ricardo and Toyota both investigated poppet-valve reverse- such a combustion system. loop-scavenged engines at a similar time [5,6], although no production applications were forthcoming. More recently Lotus Before describing the individual scavenging systems analyzed, a proposed and demonstrated a variable compression ratio (VCR) loop- short description of the Burt-McCollum sleeve valve (as used in the scavenged engine which could be started and run in homogeneous- Rolls-Royce Crecy) will be made, since these mechanisms have not charge compression ignition (HCCI) on a wide range of fuels and been used in mass-production engines for over 50 years and this will with no assistance from spark ignition [7,8], and the Achates Power aid in understanding two of the systems investigated here. have helped to promote interest in the opposed-piston engine for medium- to heavy-duty on-highway applications [9,10,11,12]. The Burt-McCollum Sleeve Valve The disadvantages of the 2-stroke cycle primarily stem from the lack of a dedicated exhaust stroke making scavenging of the burnt gases A sleeve valve is a form of gas exchange mechanism which fits and their replacement with fresh charge problematic. It is not often slideably as a sleeve between the cylinder wall and the piston. Two realized that a 4-stroke cycle engine actually swaps functions every main types of such have been productionized: the Knight other revolution of the crankshaft: half of its life it is a combustion double sleeve and the Burt-McCollum single (or mono) sleeve. engine and the other half it is a very heavily-constructed scavenge pump. A 2-stroke uses its -rod-slider mechanism as an engine The Knight double-sleeve mechanism uses two concentric sleeves continuously, but needs a means external to the which reciprocate, each driven by a crank-rod-slider-type mechanism to achieve gas exchange. This is true whether it uses the underside of at half engine speed (for the 4-stroke engines that they were the piston or a separate pump (so-called crankcase or external exclusively used in in production). Ports in the top of the sleeves scavenging arrangements respectively). enabled operation on the 4-stroke cycle due to the way in which the sleeves were timed and how the ports interacted with each other. Functionally, because the exhaust and intake phases necessarily These engines were reportedly very quiet in operation, but were overlap to a degree, there is considerable potential for charge short- costly to produce and consumed oil at a rate markedly above then- circuiting. For a simple crankcase-scavenged piston-ported engine current engines with poppet valves [16]. They were also problematic with mixture formation outside of the engine the unburned fuel loss is to start when cold: a corollary of all of the sliding surfaces they had significant. This leads to high unburned hydrocarbon (UHC) and the fact that they all merely reciprocated along a single line. For emissions in the exhaust and also to a significant increase in fuel these reasons they were not considered successful. consumption. This works counter to the theoretical advantages of the 2-stroke engine discussed above. The Burt-McCollum1 sleeve valve, while at first glance similar, is fundamentally different. Here the single sleeve moves in an elliptical Because of the likelihood of charge air loss direct from the cylinder, path, providing two dimensions in which ports in the sleeve can align achieving lambda = 1 in the exhaust is very difficult. This being the with intake and exhaust ports in a 4-stroke engine. A representation case, the use of a simple three-way catalyst to control emissions (as of a Burt-McCollum sleeve is shown in Figure 1. Conventionally the used in conventional 4-stroke SI engines) is impossible. This is the sleeve was driven in this elliptical path by a peg which was mounted reason why in this work gasoline compression ignition is used as the on a half-engine-speed gear; a hole in a drive ball, itself held in a combustion mode, as will be discussed later. socket carried in an extension to the sleeve base, allowed the peg to slide within it and the resulting motion to become elliptical as the The way in which throttling losses are reduced through the removal sleeve moved in an arc as it simultaneously reciprocated. Figure 1 of the 4-stroke engine’s positive scavenging strokes means that at shows the extension to the sleeve that the sleeve ball carrier shown in part load the 2-stroke engine necessarily operates with large Figure 2 bolted to in the case of the Bristol Centaurus engine [17]. quantities of residual gas. For SI applications this gives rise to poor combustion, which worsens as the load (and with it the amount of air flowing into the cylinder and available to displace the burnt gas) decreases. Eventually some cycles completely fail to ignite because they are below the ignitability limit, in turn meaning that the next one will have a higher proportion of fresh unburned charge, which then 1 The Burt-McCollum single sleeve valve contains elements that permits combustion initiation. The engine alternately fires and is then were separately invented by Peter Burt, a Scottish inventor and said to be ‘4-stroking’. This is not a term of endearment, since UHC designer of engines for , and James McCollum, a Canadian emissions and fuel consumption both then deteriorate further. engineer. In the form that it was put into mass production it arguably owes more to the former than the latter, and hence, possibly, the Diesel applications do not suffer in the same way but, especially for order of their names adopted in describing it [17]. loop-scavenged engines, without the structured air motion resulting Page 3 of 21

7/20/2015

Fig. 1: CAD model of Burt-McCollum sleeve as used in the Bristol Centaurus aero engine. Reproduced from [17].

Fig. 3: CAD model of the crankshaft and a cylinder set from the Bristol Centaurus aero engine which used Burt-McCollum sleeve valves for gas exchange. Reproduced from [17].

Because it had fewer parts the Burt-McCollum sleeve valve was much cheaper to produce than the Knight arrangement. Famous engines to use the single-sleeve valve include many of the Bristol air- cooled radial aero engines and the . The Centaurus, an 18-cylinder twin-row and at 53.62 litres the largest of the Bristol reciprocating aero engines, reportedly had the longest of any piston aero engine, and once external sleeve-contacting rings (which were stationary in grooves in the Fig. 2: CAD model of ball-and socket joint arrangement used in the Burt- cylinder wall and acted inwards on the base of the sleeve) were McCollum sleeve as used in the Bristol Centaurus aero engine. The sleeve drive peg is on the crank that was in turn driven by a half-engine-speed gear in adopted and the importance of oil temperature control was realized it this engine. Reproduced from [17]. reportedly had extremely low oil consumption too [18].

Figure 3 shows a representation of a cylinder set from the Bristol In addition to being desmodromic mechanisms, both Knight and Centaurus 4-stroke engine, whose design was used for prior work Burt-McCollum valves conventionally employed what amounts to a investigating the angle-area potential of a sleeve valve versus a four- fixed piston as a ; this is termed a and is poppet-valve-per-cylinder arrangement [17]. This previous study visible in Figure 3. It carries junk rings – stationary piston rings – to showed that even applying the latest limits on valve sizes and lift permit sealing of the top of the combustion chamber. Because there profiles, the single-sleeve valve offers significant extra breathing are no poppet valves, the combustion chamber can be ideally shaped potential over the current automotive state of the art. Figure 3 shows and cooled; there was ample room for the twin spark plugs necessary the sleeve towards the top of its travel, indicating that the piston is at in the aero engines that such valves were very successfully applied to, top dead centre (TDC) firing. for instance.

Sir Harry Ricardo was a major proponent of single-sleeve valves, and in addition to supporting Bristol, Napier and Rolls-Royce in their 4- stroke applications he had experimented with their application in 2- stroke engines [19]. The specific embodiment championed by Ricardo was to use the sleeve as a control to permit asymmetric timing of a uniflow scavenging system, with the exhaust port at the top and intake at the bottom. Here he took things considerably further and applied stratified charge fueling as well, with the resulting combustion system being able to operate unthrottled over large portions of the operating map. Both Rolls-Royce and Napier built

Page 4 of 21

7/20/2015 test engines but with the latter occupied in developing the Sabre engine it was Rolls-Royce that received a development contract for what became the Crecy, a 26.1 litre 90° V12 2-stroke engine intended for high-speed interceptor aircraft applications. A sectional drawing of the Crecy is shown in Figure 4 [20], and its operating cycle is depicted in Figure 5 [21] with approximate event timings after and before top dead centre (ATDC and BTDC, respectively); in Figure 5 the arrangement of the fuel and twin spark plugs in the combustion chamber bulb can readily be discerned.

Fig. 5: Rolls-Royce Crecy operating cycle. 1: combustion (0° ATDC); 2: end of expansion with exhaust port at top of sleeve valve about to open (start of blowdown) (90° ATDC); 3: exhaust and intake ports open – note upwards uniflow scavenging flow (135° ATDC); 4: exhaust ports starting to close (end of scavenging), start of (180° ATDC); 5: end of fuel injection, supercharging via intake ports (°135 BTDC); 6; start of Fig. 4: Sectional drawing of the Rolls-Royce Crecy sleeve-valve 2-stroke compression, with rich mixture trapped in combustion chamber bulb (120° engine (in its original non-compounded form). Reproduced from [20], BTDC). Reproduced by permission of Rolls-Royce Heritage Trust from [21], courtesy of The Rolls-Royce Heritage Trust; copyright the estate of Lyndon copyright Rolls-Royce plc. Jones. Inspection of Figures 4 and 5 readily shows that the Crecy employed an elegant form of sleeve drive via yoke plates driven directly by eccentrics on the crankshaft (this being possible since it was a 2- stroke engine). These had the drive pegs attached to them, and each yoke drove two sleeves. A dummy piston was used to stop yoke rotation. This mechanism still gave rise to an elliptical motion to the sleeve. A final mechanical point of interest is that while the intake ports in the sleeve were necessarily composed of windows, the exhaust was of a ‘360°’ configuration, i.e. it was pulled clear of the junk head in its entirety with no windows. This gave the minimum duration for the necessary angle-area but also meant that blowby into the void above the sleeve and then out into the exhaust port would have been considerable. This will be returned to later.

A representation of the port timing is given in Figure 6; this is from work conducted for this paper in order to give angle-area information for the one-dimensional (1-D) analysis work conducted here. This graphical construction was undertaken from information given in [21].

Page 5 of 21

7/20/2015 4. The reverse-uniflow sleeve-valve engine. This is based on the Crecy arrangement and is referred to here as the ‘reverse- uniflow sleeve’. Conceptually it might be considered to be a hybrid of the Crecy and Clerk’s engine of 1881. It was analyzed partly due to the findings of an earlier paper which compared it to the OP2S and port-poppet engines [34].

5. The reverse-loop poppet-valve engine, as previously proposed by Toyota and Ricardo [5,6] and currently being investigated by Renault and others [14]. This shares the most clear architectural links with the current automotive norm, the poppet-valve 4-stroke engine. This type was introduced to give a more complete comparison using technology in production now.

6. The Schnürle-loop scavenged engine, referred to in this work as ‘loop-scavenged’2. This is thought of by many as the classical Fig. 6: Example of port timing drawing development undertaken for the Rolls- 2-stroke , in the same way that the poppet- Royce Crecy for the work reported here. valve 4-stroke engine is now considered a norm [1,35].

Finally, after World War 2 the sleeve-valve 2-stroke engine was These arrangements are shown schematically in Figures 7 to 12. further investigated by S.S. Tresilian for Rolls-Royce as a major component in a highly-integrated aircraft propulsion system. This was the so-called ‘X-engine’ [21] which sought to maximize the architectural possibilities of the sleeve valve within a 16-cylinder 4- row in-line radial engine, ultimately to have been of approximately 9.0 litres swept volume and employing high levels of turbocompounding. It did not progress as a concept, because of the rapid improvement of the occurring at that time. The subject of turbocompounding, and which of the different engine types discussed in the present work might suit it best, is returned to later.

For more detailed information on sleeve valves and their application see [20,21,22,23,24,25].

Scavenging systems investigated Fig. 7: Schematic representation of the opposed-piston engine (OP2S). In this diagram the exhaust piston and ports are at the top and the inlet piston and The work presented here seeks to compare six different scavenging ports are at the bottom. Port timing is controlled by the , which do not systems for a 2-stroke engine suitable to power a US light-duty truck. have to be in phase, making asymmetric port timing possible. The primary The scavenging system effectively defines the major architecture of a scavenging air motion is upwards. 2-stroke engine, and there are several fundamental types that can be utilized. Together with the combustion system it dictates the performance and fuel consumption of the engine. Sher provides an excellent overview of some scavenging types and their history [1]. The way in which the comparison was made is described later. The scavenging systems investigated here were:

1. The opposed-piston 2-stroke (OP2S) engine, which has successfully been applied to aircraft propulsion as well as to engines for power generation and rail traction and has recently been promoted by Achates Power for heavy-duty applications [9,10,11,12,20,26,27,28]. Fig. 8: Schematic representation of the poppet-valve uniflow configuration. In this diagram the exhaust valves and ports are at the top and the inlet ports 2. The poppet-valve uniflow configuration, as exemplified by the Detroit Diesel 2-stroke engine [29] and various marine engines [30,31,32] and which is being reevaluated now for automotive use [33]. This is referred to herein as the ‘port-poppet’ type. 2 Sher [1] describes the true Schnürle-loop scavenging as being only by intake ports to the sides of the exhaust port. He refers to the 3. The sleeve-valve uniflow arrangement, which was used in the evolved configuration which most now consider to be Schnürle-loop, Rolls-Royce Crecy described in the previous section. In this having intake ports opposite the exhaust port as well, as being work this will be referred to as the ‘forward-uniflow sleeve’ ‘Curtis-type’. Here the term Schnürle-loop will be used exclusively [21]. to encompass this evolution as well.

Page 6 of 21

7/20/2015 are at the bottom and uncovered by the piston. Asymmetric timing is combustion chamber and down, with this being assisted heavily by the ports’ obviously possible. The primary scavenging air motion is upwards. angles to the cylinder.

The Schnürle-loop scavenged engine has historically had various complications applied to it to improve its efficiency, with a valve in the exhaust similar to the ‘YPVS’ system used by Yamaha being very common. Such valves vary the exhaust port height non-cyclically, and so affect exhaust port opening (IPO) and closing (IPC) timing equally. Since the Schnürle-loop scavenged configuration represents what many consider to be the conventionally-conceived 2-stroke engine, it represents something of a baseline for the other arrangements. It was not assessed as part of the previously-reported Fig. 9: Schematic representation of the forward-uniflow sleeve-valve work [34] and so it was included here. However, in including it it arrangement, as used in the Rolls-Royce Crecy. In this diagram the exhaust was also decided to assess the most flexible of the exhaust timing ports are at the top and the inlet ports are at the bottom. Both sets of ports are adjustment mechanisms applied to this type, which was considered to uncovered by the sleeve which moves elliptically. Asymmetric timing is be the Lotus charge trapping valve (CTV) described by Blundell and possible. The primary scavenging air motion is upwards. co-workers in several engine research projects [7,8,15,36]. This would set a ‘best case’ baseline to compare the other scavenging systems to3.

Configurations (1), (2) and (3) were described in a previous paper in which only uniflow scavenging arrangements were compared [34]. In that work the OP2S engine emerged as a clear winner, with the port-poppet configuration second but only marginally better than the forward-uniflow type. In [34] the observation was made that the port-poppet type was severely restricted by valve train dynamics and as a consequence, since it is desmodromic and is not limited by such Fig. 10: Schematic representation of the reverse-uniflow sleeve-valve issues, it may well be that the forward-uniflow-sleeve might be scavenging arrangement. In this diagram the inlet ports are at the top, and the superior under some situations. In the analysis of the angle-areas exhaust ports are at the bottom. The ports are uncovered by the sleeve which undertaken for the Crecy for the earlier work, it was determined that moves elliptically. Asymmetric timing is possible. The primary scavenging the flexibility provided by the sleeve’s port layout could permit air motion is downwards. additional optimization possibilities. Furthermore, with the scavenge air flowing in the reverse direction (i.e., from the top down) in this new layout, the combustion chamber and the top of the sleeve could be expected to run cooler, although it was accepted that the exhaust leaving the base of the cylinder would likely mean that the piston would then run hotter (a common challenge for the exhaust piston of the OP2S configuration as well). A cooler-running combustion chamber might be expected to extend the knock limit in SI applications. As a consequence, configuration (4) has been included in the present work to assess whether it has merits in comparison to the OP2S.

Fig. 11: Schematic representation of the reverse-loop-scavenged poppet-valve All of the concepts were compared in terms of net indicated fuel engine. Intake and exhaust valves are situated at the top of the cylinder as per consumption for the same cylinder swept volume of 751 cc, being an the current automotive norm. In this configuration asymmetric timing is individual cylinder swept volume suitable for a medium-to-heavy possible using conventional phasing devices. The primary scavenging air duty engine (i.e. the sectors where it is expected that such high motion is down, across the piston crown and up, with this being heavily efficiency 2-stroke engines will be introduced first). Similarly a assisted by the ports’ angles to the cylinder; the intake is conventionally geometric compression ratio (CR) of 15:1 was adopted for each. vertical and situated between the . However, since the ratio of stroke length to diameter cannot be kept constant for all of the arrangements without severely compromising one or more of them, this was altered to better suit their requirements; for example, the OP2S engine used the same total stroke:bore ratio of 2.2 as Achates [11,12]. Once these were set, no further individual optimization of this variable was conducted. It was

3 Schnürle-loop scavenging represents a clear advantage over earlier piston-ported single-piston designs such as cross- or loop-scavenging Fig. 12: Schematic representation of the Schnürle-loop scavenged engine. [1]. The latter are rarely used now. The required deflector piston Both the exhaust and intake ports are positioned at the base of the cylinder and detail for cross scavenging compromises the combustion chamber their primary timing is controlled by the piston. In this configuration shape, and in the conventional loop-scavenged engine the position of asymmetric port timing is not possible without further complication being the exhaust port above the transfer means charge trapping is poor. added (see text). The primary scavenging air motion is up, across the Page 7 of 21

7/20/2015 also assumed that a 4-stroke-style wet crankcase design would be Three engine operating points were used to compare the different used, i.e. that all engines would use external scavenging (see below). scavenging system designs. These were: Operationally, the gasoline compression ignition (GCI) combustion system that was modelled for all variants used the same combustion A 1500 rpm, 3 bar IMEP, 1.2 bar pressure profile for each of the operating points investigated (see later), this being the result of previous work conducted by Aramco [37]. This B 1500 rpm, 14 bar IMEP, 2.0 bar exhaust manifold pressure combustion system was adopted both to simplify the modelling process and because it provides very low engine-out NOx, which will C 3000 rpm, 12 bar IMEP, 2.5 bar exhaust manifold pressure be of crucial importance to 2-stroke operation where it is difficult to provide the lambda = 1 conditions in the exhaust necessary for a three-way catalyst to work properly. These operating points took into account the medium-duty nature of the study: point A was intended to be a representative part-load operating point, and B and C were notionally peak torque and peak The engine specifications used in this work are listed in Table 1. power respectively; they will be referred to in this manner going forward. Points B and C turn represent nominal indicated specific Table 1. Engine specifications modelled. outputs of 225 Nm/l and 60 kW/l for this cylinder capacity, i.e. reasonable performance for this type of application. Engine Opposed- Port- Sleeve- Sleeve- Poppet- Piston- Type Piston Valve Valve Ported (OP2S) Forward Reverse Reverse Loop Each model was created as a generic single-cylinder version of the Uniflow Uniflow Loop (Schnürle) specific concept using GT-Power, a 1-D engine simulation software package. The models consisted of the cylinder and porting Basic Uniflow Uniflow Uniflow Uniflow Loop Loop Scavenging arrangements, with user-imposed conditions either side via a series of System 0-D elements. The OP2S engine was modelled as an equivalent single cylinder where the piston area was doubled and cylinder head Bore [mm] 75.75 86 86 86 98.52 98.52 area set to zero to represent the opposed piston arrangement. The Stroke [mm] 166.65 129.29 129.29 129.29 98.52 98.52 equivalent single piston motion was calculated to account for the relative motion of the two pistons. For all models, the direct fuel Stroke:Bore 2.2 1.5 1.5 1.5 1 1 injection quantity was calculated from the desired air-fuel ratio Ratio [-] (AFR). At operating point 1 the AFR was held at 43.7 for all configurations, while at operating points 2 and 3 the AFR was Conrod 166.65 258.58 258.58 258.58 197.04 197.04 Length increased to 16.2 for all cases. As mentioned above, these were the [mm] result of earlier work by Aramco on GCI [37]; the heat release profile adopted is shown in Figure 13. Cylinder +4.56 0 0 0 -1.73 -1.73 Surface Area difference [%]

Note that for the OP2S engine, the terms TDC and bottom dead centre (BDC) are referenced from the exhaust piston angular position; typically the exhaust piston leads the intake.

Before describing the modelling methodology adopted, it should be noted that no attempt was made to specify and model an equivalent 4- stroke engine. This was due to the complexity of the undertaking and its difference to the majority of the engines. This was considered Fig. 13: Combustion heat release profile adopted from work performed by acceptable because other researchers have already done this for some Aramco on GCI [37]. scavenging configurations. Specifically, in [38] it was estimated that a OP2S would have 13-15% better fuel consumption Intake air pressure was controlled to achieve the target IMEP via a than its 4-stroke equivalent, and later work by those authors closed loop. While the exhaust pressure was user-imposed, exhaust suggested that a OP2S would still have 7% improvement and be pressure sweeps were also performed to verify trends that were seen broadly equivalent in fuel consumption terms to a compounded 4- at the individual operating points. stroke engine [39]. This will be returned to later. To minimize friction and achieve a degree of waste heat recovery Description of the simulation method scavenging air supply was assumed to be provided by a system comprising turbocharger and supercharger in series, although this In this paper we focus on the simulation results for lowest fuel was not explicitly modelled. In order to evaluate fuel consumption consumption, and all of the results quoted are given on an indicated meaningfully, the power to drive the mechanical supercharger had to basis. This is because the construction of full engine models and the be accounted for, and so a net specific fuel consumption (NSFC) was estimation of friction losses was outside the scope of the current level calculated as follows: of work; however it is recognized that sleeve valves would likely have different friction and heat loss behaviours. 푁푆퐹퐶 [𝑔⁄푘푊ℎ] =

Page 8 of 21

7/20/2015 퐹푢푒푙 퐹푙표푤 [𝑔⁄ℎ] 3. Scavenge-specific time area – the minimum of the exhaust and

(퐼푛푑푐푎푡푒푑 푃표푤푒푟 [푘푊] + 푆푢푝푒푟푐ℎ푎푟𝑔푒푟 퐷푟𝑖푣푒 푃표푤푒푟 [푘푊]) intake port open areas over the interval EPO to exhaust port closing (EPC) Eqn 2 The guidelines recommended by Naitoh and Nomura were used to Supercharger drive power was estimated from an energy balance of establish angle-area targets to lead the design of the port geometry at the exhaust and intake conditions, with the difference between what the beginning of the study [43]. These were originally derived for was required and what the turbocharger could supply assumed to be high-performance Schnürle loop-scavenged engines for motorcycle provided by the supercharger. In order to remove the effect of racing, but nevertheless it was believed by the authors that they changing assumptions, all turbomachinery efficiencies were assumed should applicable to any 2-stroke scavenging configuration, and that to be 70% in this work. they could be used to get sufficiently close to the eventual configuration that a numerical optimizer could then be used (see later). It is believed that this is the first time such an approach has Since no detailed in-cylinder flow modelling was conducted, the been taken across such a broad range of engine layouts. Although scavenging behavior of each design was dictated by a profile which not tested here, it is further believed that this should be the case relates in-cylinder to exhaust manifold burned gas fractions. This regardless of whether it crankcase scavenging or an external was determined from an extensive survey of the literature and scavenge pump is employed. changes between scavenging types. These profiles are shown in Figure 14; the scavenge profile for the OP2S was taken from work by Mattarelli et al. [40], for the port-poppet and both sleeve valve The general constraints on the engine operating envelope for arrangements from a uniflow study described by Laget et al. [41], optimization were logically set to be that under normal operation and for the reverse-loop poppet-valve it was adapted from the EPO should be before IPO and EPC should be before IPC (i.e. Renault Powerful concept [42]. asymmetric timing). As discussed above it was assumed that to supply air to the engine some form of compound charging system would be required, and the work necessary to drive the supercharger component was calculated and applied so that this requirement was accounted for. To reflect this a further constraint was imposed, so as to ensure that there would be sufficient exhaust pressure available to drive a turbocharger in such a system: the pressure in the volume immediately downstream of the exhaust port was set to 1.2 bar at the part-load operating point, to 2.0 bar at peak torque, and to 2.5 bar at peak power (points A, B, and C respectively). These values in turn assume that a meaningful exhaust back pressure downstream of the turbine for a catalyst and silencer system has been catered for.

In terms of process, port timings were determined by numerical optimization of the models at the 1500 rpm 14 bar IMEP maximum torque operating point (Point B)4. This was because that while establishing a ranking for minimum in-use fuel consumption was the primary aim, the engines had to meet the performance targets to be viable, and it was felt that the step in air mass flow from peak torque Fig. 14: Scavenge profiles used in the models [40,41,42], together with other to peak power could be accommodated by changing the boost example profiles taken from literature. pressure. NSFC was minimized at this point within the constraints imposed by the geometry and design of each concept. These port When analyzing the port and valve optimization results specific time- timings were then applied at the other two operating points. area was used extensively. This metric has historically been used to assist in the design of 2-stroke engine ports and was used here as a guide. The specific time area is defined to be the integral of port Engine porting arrangements open area with time divided by the swept volume. It provides a measure of port availability for gas flow during the cycle and there The general port timing arrangements for the OP2S engine are are different values which therefore affect the performance of an controlled by their respective pistons; the lead of the exhaust over the engine: intake is an important parameter because it is what gives it its asymmetric port timing. A study was conducted to investigate the 1. Intake-specific time area – the intake open area calculated over optimum amount of exhaust piston angular lead over the intake. the time interval from intake port opening (IPO) to intake port While it is theoretically possible to vary the timing between the two closing (IPC) crankshafts, this was not used as a variable in the present study; once the exhaust lead had been set for Point B it was left fixed for the 2. Blowdown-specific time area – the exhaust open area other operating points investigated. calculated over the time interval from exhaust port opening (EPO) to IPO (this is sometimes referred to as the free exhaust period)

As well as the above established metrics, an additional one was used 4 The term ‘port’ is used in reference to timing events throughout this to help gain insight into the scavenging process: paper even when valves and not pistons are used for this purpose.

Page 9 of 21

7/20/2015 During the exhaust lead optimization process the injection timing for similar approach to that used for the OP2S engine; a set of junk rings the OP2S was also compensated to allow for the fact that as the was assumed to be included to seal the top of the combustion piston phase changes, so the angles of maximum and minimum chamber. volume also change, effectively varying the position of TDC and BDC as far as the engine cycle is concerned (i.e. the minimum and At the other end of the cylinder for the forward-uniflow sleeve-valve, maximum volume in the cylinder, respectively). Regardless of this, the sleeve controls intake angle-area and timing. It can be timed with the swept volume was held constant at 751 cc and the geometric CR a lead or lag angle relative to the piston. These were another set of at 15:1. variables that needed to be optimized.

In his 1996 book Blair recommended that a maximum port width of Similar methodologies were applied to both the forward- and reverse- 66% could be used to avoid the use of pegged rings [35]. It being uniflow sleeve-valve engine configurations, but changes in heat expected that improvements in materials might permit an increase, transfer were not considered for the two sleeve valve configurations 75% port width open area was adopted here. This was justified for two reasons: firstly this was primarily a study on gas exchange, because if necessary the rings could be pegged anyway. For a total of and secondly there is conflicting evidence regarding heat transfer 12 ports, this metric gave 22.5° for the subtended angle of each port. being better and not worse for sleeve valve engines [19,21]. Ma et al. covered the topic of port widths comprehensively for intake Supposedly the beneficial case arises because while it represents an ports on a port-poppet engine recently [33]. For fairness of additional barrier, the elliptical nature of the sleeve motion efficiently comparison, this value was adopted for all of configurations using moves the heat around. During this work it was not considered that piston-controlled ports, including the sleeve ports for the sleeve-valve sufficient knowledge was available to influence calculations and thus engines. However, note that in those cases the respective ports could the heat transfer was considered to be similar to that of the port- be larger within the cylinder wall itself since the rings only rub poppet arrangement for ease of comparison. Further work would be against the internal diameter of the sleeve. useful to assess this.

Having set these parameters the main variable left for the cylinder For the reverse-loop scavenged engine the valve train kinematics ports was their height. This value affects angle-area and the port limitations discussed above for the port-poppet engine obviously also timing of the engine simultaneously, and thus the trapped expansion apply. However, whereas the port-poppet engine can use its four ratio (ER) and the ratio between these two (see below). To maximize valves entirely for exhaust angle area, the reverse-loop engine cannot, work and minimize NSFC one requires maximum ER together with and its breathing capacity is severely handicapped as a result. Thus the highest value of the ratio of ER to CR; theoretically if this value there are two significant limitations for this type of engine: reaches greater than unity then one can create a degree of Miller kinematics and breathing. To these must be added that charge short- cycle operation. Because of the temptation to perceive a 2-stroke circuiting is also problematic with the exhaust valves/ports engine as a symmetrically-timed piston-ported device, this is an juxtaposed with the (unlike in the piston-ported loop- operating regime that is not normally associated with this type of scavenged engine, where they are on the bore periphery), such as in engine. the ‘Flagship’ engine proposed by Ricardo [5,6]. Placing the intake ports vertically between the camshafts forces those components apart, While the intake port geometry approach for the port-poppet engine in turn making the combustion chamber more like a traditional pent is the same as that for the OP2S, its exhaust process is controlled by roof, but potentially further exacerbating the short-circuiting problem cam-driven valves. Since the intention was to maximize expansion since the ports are then angled towards each other. Ricardo reported then logically the greatest amount of valve curtain area possible compromises with respect to valve sizing, kinematics and scavenging would be needed and so in this study four exhaust valves were used [6], with a suspicion that a severe preignition problem may have been for this type. The cam profiles were calculated for typical 4-stroke one result of the adoption of the architecture. Benajes et al. reported exhaust valve reciprocating masses and the spring rates selected from a different layout (i.e. non-pentroof) for a reverse-loop scavenged an existing 1-D engine simulation model. Scaling rules were used engine [14], with promising results, but they noted that adopting an modify them to ensure that valve accelerations and velocities were early exhaust valve opening (EVO) timing for scavenging purposes not exceeded for use in the 2-stroke engine, i.e. that valve control reduced expansion and compromised fuel consumption, echoing the would be maintained over the engine operating range investigated, results found here. On the positive side for this configuration, twin accommodating the fact that the camshafts are now turning at engine phasing devices could readily be adopted to tune intake and speed. As discussed in previous work on such engines, the exhaust port timing relative to piston motion (assuming the required limitations on port angle-area imposed by valve kinematics are a operation of the at crankshaft speed was not a limitation). factor which limits its performance [34]. This is essentially because the kinematics set a minimum valve event length and the resulting After the initial shake down of the models the severe limitations on process has to be timed to have the minimum impact on the trapped breathing with this type of engine meant that its performance was CR and ER, in turn limiting work extraction as discussed above. sufficiently poor in relation to the other types that, after due consideration, it was not taken further to a full optimization. Hence it For the forward-uniflow sleeve-valve engine a general engine is not included in the discussion section that follows. However, this cylinder scheme was drawn using the dimensions of Crecy. This was is not to say that as a concept it is entirely without attraction: the then scaled appropriately for the engine being modelled. In Figures 5 ability to share machining lines with existing production 4-stroke and 9 it can be seen that the exhaust exits at the top of the cylinder engines and the fact that piston rings do not have to traverse ports at like the port-poppet engine, but with 360° of port width. While this all (unlike with all of the other configurations discussed here) are approach logically minimizes port height for any required angle-area, considered to be two major ones, as originally discussed by for the application considered here this was considered impractical researchers at Ricardo [5,6]. Nevertheless, in the present research, because crevice volumes have to be controlled and pressure loss which was aimed at maximizing efficiency and for which maximum contained, both of which were not considered serious issues in the expansion ratio is required, the fact that valve train kinematics force Crecy. Hence the sleeve was modelled with lands and angles using a Page 10 of 21

7/20/2015 long events which then compromise this aspect of performance was ultimately what led to its being deselected at this point.

Finally, the loop-scavenged piston-ported engine was based on information in the published Lotus 2-stroke papers where a CTV had been employed to modify and tune the exhaust event [7,8,15,36]. The CTV permits asymmetric timing of the exhaust port and can also be variable to permit optimization. The ability to adopt the publicly- available data of Blundell and co-workers for the models was one reason why the stroke-to-bore ratio of this engine was fixed at 1.0, and why it was decided to adopt different values of this parameter for Calculate initial IVO, IVC, EVO, and different engine architectures. EVC using specific time-area guidelines from Naitoh and Nomura Once these engine porting arrangements had been decided upon, a [43] and the geometric constraints of staged approach was taken for the optimization process for the port each design timing and geometry:

1. The engine models were run at the operating point B (peak Sweep IVO and EVO at operating point torque). Sweeps of the intake and exhaust port timings were B by back-calculating to geometric then performed. parameters (e.g. port height w.r.t. BDC)

Surface response models of the variables of interest (primarily NSFC) were created as functions of the intake and exhaust Create surface response models from timing events. the sweep data for NSFC, ISFC, and supercharger work 2. An offline optimizer was then applied to these response models to find the optimum whilst adhering to the constraints based on geometry and operating conditions discussed above. The Use optimization algorithm to locate resulting optimal timings were then used to calculate the IVO / EVO timings that minimize NSFC port/valve geometries required to achieve them.

A schematic diagram of the process followed is shown in Figure 15.

Fig. 15: Schematic diagram of the process of the simulation method used to determine port geometries and timings.

The resulting port geometries and timings were then used to predict the performance at operating points A (part load) and C (maximum power). The NSFC results were then averaged in order to rank the scavenging systems.

Results

The results of the optimization are summarized in Table 2, and the associated port profiles are shown in Figures 16 to 20. Table 3 presents the port timings and Table 4 the associated specific time- area values. For the reasons discussed above, the reverse-loop poppet-valve configuration is not included in any of the results due the indications from early in the project that it would clearly be inferior to the other configurations, meaning that it was not taken forward into the optimization stage.

It is important to note that since the engines have not been modelled in detail the data shown should not be interpreted as absolute values

Page 11 of 21

7/20/2015 for them, although it can be used for a general comparative ranking of the different approaches.

Table 2: Comparison of results for the different scavenging concepts studied.

Engine Type NSFC [g/kWh] Estimated Supercharger Power Requirement [kW]

Operating A B C A B C Point

OP2S 183 189 192 0.244 0.158 1.96

Port-Poppet 194 211 207 0.210 1.27 3.04

Forward- 197 214 208 0.224 1.27 3.01 Uniflow Sleeve Fig. 18: Port area profiles optimized for lowest NSFC at operating point B (peak torque) for the forward-uniflow sleeve-valve engine (Rolls-Royce Crecy-type). Reverse- 187 214 201 0.197 2.02 2.68 Uniflow Sleeve

Piston-Ported 199 214 211 0.230 0.956 3.18 Loop

Fig. 19: Port area profiles optimized for lowest NSFC at operating point B (peak torque) for the reverse-uniflow sleeve-valve engine.

Fig. 16: Port area profiles optimized for lowest NSFC at operating point B (peak torque) for the OP2S engine.

Fig. 20: Port area profiles optimized for lowest NSFC at operating point B (peak torque) for the Scnürle loop-scavenged piston-ported engine (utilizing a Lotus-type charge trapping valve to achieve asymmetric port timing).

Table 3: Numerical summary of port timings of the different concepts.

Engine Type Optimized Valve / Port Timings [°ATDC]

Fig. 17: Port area profiles optimized for lowest NSFC at operating point B (peak torque) for the port-poppet uniflow engine. EPO EPC IPO IPC

Page 12 of 21

7/20/2015 OP2S 140 220 147 228 Average of Change Average of Change three relative to three relative to operating Loop- operating Loop- points Scavenged points [kW] Scavenged Port-Poppet 115 225 145 215 [g/kWh] Piston-Ported Piston-Ported [%] [%]

Forward- 113 218 145 216 OP2S 188 -9.6 0.79 -45.9 Uniflow Sleeve

Port-Poppet 204 -1.9 1.51 +3.4 Reverse- 137 239 149 252 Uniflow Sleeve Forward- 206 -1.0 1.50 +2.7 Uniflow Piston-Ported 113 220 140 220 Sleeve Loop

Reverse- 201 -3.4 1.63 +11.6 Uniflow Sleeve Table 4: Numerical summary of specific time-areas of the different concepts. Piston-Ported 208 - 1.46 - Loop Engine Type Specific Time Areas (all at 1500 rpm) [s.cm2/cm3]

Intake Blowdown Scavenge Finally, while the reasons for the results of the individual concepts OP2S 11.8E-05 2.78E-06 9.99E-05 are discussed in detail below, with regards to the discussion above on the possibility of applying Miller-type operation to a 2-stroke engine, Table 6 presents a numerical summary of the CRs and ERs and the Port-Poppet 12.6E-05 7.12E-06 6.63E-05 ratios between them.

Table 6 Numerical summary of the effective (or trapped) compression and Forward- 11.9E-05 22.3E-06 6.86E-05 expansion ratios of the different concepts, and the ratios between them. Uniflow Sleeve Engine Type Effective Effective Ratio of Expansion Compression Expansion Ratio to Compression Reverse- 10.4E-05 4.74E-06 9.87E-05 Ratio [:1] [:1] Ratios [-] Uniflow Sleeve Volume at start of Volume at end of Expansion ratio / compression / expansion / compression ratio Piston-Ported 10.9E-05 15.2E-06 7.4E-05 clearance volume clearance volume Loop

OP2S 13.68 13.67 1.00

Port-Poppet 13.79 11.18 0.81 From Table 2 it is readily apparent that the OP2S engine is the best configuration with regards to fuel consumption: it has the lowest Forward- 13.85 11.49 0.83 values of NSFC for all three operating points, and indeed is the only Uniflow one below 200 g/kWh at points B and C. It also has the lowest Sleeve estimated supercharger power requirement. To reinforce this the data Reverse- 10.97 13.53 1.23 of Table 2 is averaged in Table 5 and then compared to the loop- Uniflow scavenged engine (the worst performer) in terms of percentage Sleeve change. This is done for both NSFC and the estimated supercharger power requirement, but it must be noted that this value reflects only Piston-Ported 13.73 11.49 0.84 the ‘make-up’ power (per cylinder) that a supercharger would have to Loop supply as part of a compound charging system within the modelling assumptions discussed earlier.

Table 5: Comparison of results for the different scavenging concepts studied Here it can be seen that the reverse-uniflow sleeve-valve engine has in terms of averages and change relative to the loop-scavenged piston-ported the best value of the CR/ER metric, meaning that it is effectively an engine. engine operating on the Miller cycle. Of the others, the OP2S is closest, with all of the other three scavenging systems then being Engine Type NSFC Estimated Supercharger Power Requirement approximately the same.

Figure 21 presents a breakdown of the power flow from the cylinder at the point B condition (or peak torque), which is useful for a visual comparison of pumping work and heat losses. However, please note Page 13 of 21

7/20/2015 that while the conditions are optimized for lowest NSFC at this point, loss to heat transfer is the lowest; this was the case for all of the this is not an efficiency breakdown. operating points investigated here.

The latter point is in line with what other researchers have discussed [12] but here attention is drawn to the fact that the reason is a net gain from the summation of the interaction of surface areas, heat transfer coefficients (HTCs) and surface temperatures rather than being merely a simple observation that a cylinder head is not present like in other types of engines. It is believed that this advantage could be further improved by optimization of the stroke:bore ratio. Nonetheless, as it stands, the reduced power rejected to coolant would be expected to give a benefit at the vehicle level as well.

For the OP2S the optimizer chose an IPO very shortly after EPO, resulting in the smallest exhaust blowdown period of all the concepts. However, note that the subsequent scavenge period (in terms of scavenge time-area) is the highest. It is thought that this, combined Fig. 21: Breakdown of power flow at operating point B (peak torque – 1500 with a slightly more favourable scavenge profile, results in similar rpm, 14 bar IMEP and 2.0 bar exhaust back pressure). (Note that this is not an levels of trapped residual gases to the other concepts. efficiency breakdown – see text.) In the model, the short blowdown period results in a flow of residuals The fact that from Table 2 the work conducted here shows the OP2S into the intake system at IPO due to the cylinder pressure conditions. engine to be the best of the options modelled is considered to be To some extent this is observed in some of the other designs due to some validation of the previous work by Warey et al. [38,39] in optimization of port timings for minimum NSFC, however it is most which in terms of fuel consumption the OP2S configuration was pronounced in the OP2S. found to be better than loop scavenging and even better still than the comparable 4-stroke engines that they modelled; consequently the This concept has the lowest supercharger work which undoubtedly present work is also considered to support the work of Achates Power contributes to the low NSFC. Further retardation of timing events in terms of diesel OP2S fuel consumption advantages over current 4- does reduce ISFC but then requires higher supercharger work since stroke engines. the incoming air has to be compressed to a greater pressure, and thus the NSFC increases. Discussion Given the simplified nature of the charging system model (which This section will discuss each case individually before drawing broad relies heavily on estimated efficiencies) it is not possible to say comparisons between the systems. whether or not the supercharger could definitely be de-clutched at any of the engine operating points. However, at peak torque the estimated charging system load is very small and allowing for an Opposed-Piston (OP2S) optimized turbocharger match it is possible that the supercharger could be disengaged fully, further improving fuel consumption. From Table 2 it can be seen that the OP2S engine delivers the lowest NSFC over all three selected operating points, and from Table 5 it is The optimum phase lag between the pistons was found to be an evident that it gives easily the lowest NSFC, being 9.6% lower than exhaust lead of 7.5°. While is this value is specific to the geometry the piston-ported loop-scavenged engine. This result is due to a modelled it is not very different to the 8° value settled upon for one combination of the lowest ISFC coupled to the lowest supercharger of the Achates Power engines [10]. After the main optimization work; this was 45.9% lower than the piston-ported engine. This is phase of the project had been completed, the phase offset between the achieved through two principal routes: pistons was investigated in more detail. This was done in order to investigate what effect a theoretical device that could vary the phase 1. Maximized expansion work. The mechanical arrangement between the crankshafts might have on economy. Consequently for permits later EPO which maximizes expansion work. In Table this investigation the port geometry was fixed and the CR was 6 it can be seen that the ER to CR ratio is 1.00, compared to allowed to vary with the change in exhaust piston lead. The only values of less than 1 for all of the other concepts except the parameter that did vary during this investigation was the injection reverse-uniflow sleeve-valve (i.e. they operate under- timing in order to maintain the same SOI relative to the effective expanded). This design therefore has this thermodynamic volume-based TDC, i.e. all of the hardware was essentially fixed. advantage over most of the other configurations. The results of this sweep are shown in Figure 22 for the three 2. Reduced heat transfer. Although the specific design rules different loads investigated. There was found to be small benefit in applied result in a larger total surface area compared to the ISFC and NSFC from varying piston phase for the part-load and full- port-poppet and sleeve design (approximately 4.6% greater: see power operating points, but overall the general response is Table 1), the average temperature of the combustion chamber remarkably flat, especially between 2.5° and 12.5° of exhaust lead. surface is higher (because the two pistons run hotter than a In reality, in terms of combustion control a potentially important cylinder head would). As a consequence heat transfer is advantage of such a complication would be the ability to provide reduced. The advantage that the OP2S has in this area is VCR to control GCI combustion more directly. The significant clearly shown in Figure 21, where the magnitude of the power potential of VCR in this context has been demonstrated by Turner et

Page 14 of 21

7/20/2015 al. [7,8]. It is considered that this possibility is worthy of further If there was a way to improve the poppet valve performance (i.e. investigation. shorten the duration whilst maintaining lift), it may be possible increase expansion work and improve ISFC. Such mechanisms may include the use of operation, as is used in production by Ducati, or an air-valve spring system. The latter is essentially a motorsport-only system and so is not considered viable here. Other valving systems may offer benefit, but except for the sleeve valve, these are outside the scope of this investigation.

However, when using what amounts to a conventional production valve system, there is the scope to employ variable (VVT) afforded by camshaft phasing devices. When investigating this the results showed that with nominal timing optimized for peak torque, further retardation of EVO causes a small reduction in NSFC for the part load and maximum power operating points. At the peak torque condition, retarding the timing reduces residuals, probably due to increasing the scavenge time-area, although this again increases NSFC due to higher supercharger work. It is thought that using a reverse-uniflow configuration might have some potential benefit, since applying VVT to the intake instead might then facilitate the application of Miller-cycle operation at certain operating points.

Forward-Uniflow Sleeve Valve

Because it is a fundamentally different and desmodromic mechanism, the valve kinematics issues with the port-poppet arrangement can be bypassed by the sleeve valve. This, coupled to the fact that the ports can be disposed around the cylinder periphery, means that like the OP2S the sleeve valve can provide short durations for any given angle-area requirement. However, the interaction of sleeve, piston,

and port makes its optimization more difficult due to geometric Fig. 22: Effect of exhaust piston lead angle (or piston phase) on ISFC, NSFC, constraints: for example, varying the phase of the sleeve motion burned residual rate and trapping ratio for the OP2S engine. relative to the piston changes the exhaust port timing at the top of the cylinder, whereas the intake timing is essentially piston-controlled The reader may like to note that both the optimized value of exhaust via an interaction with the sleeve ports at the bottom. This effect can lead obtained here and those of the Achates engines are significantly be seen for the Crecy in Figure 5. different to the value adopted by the historically-important Napier Deltic OP2S engine [44]. That engine utilized an exhaust lead of 20° The data of Table 3 shows that despite this extra degree of freedom, not because of performance considerations but because its unusual the optimizer converged on a set of timings very similar to the port- architecture (of three crankshafts and three banks of cylinders) poppet design. As a consequence the resulting simulated NSFC and required it to work from a mechanical and firing interval point of estimated supercharger work requirement is also very similar for view. This limitation was not shared with all other multi-bank OP2S these two concepts, as shown in Tables 2 and 5. Table 6 shows that engine such as the four-bank (and four-crankshaft) Junkers Jumo 223 the trapped CRs and ERs are very similar for the two, as well. This and 224 engines [27,28,44]. concept has the largest blowdown specific time-area of all the designs, however it has very similar scavenge time-area to the port- Port-Poppet poppet engine. The configuration was also considered interesting from the results The average NSFC is higher for the port-poppet engine versus the gathered because it had high potential for breathing at higher engine OP2S, but still 1.9% lower than the piston-ported engine. It is speeds. The limiting factor in this was found to be the interaction of constrained by acceleration forces in its valvetrain. The profiles used the piston and sleeve motions near to BDC. This was one of the here were taken from a model of a modern 2.0 litre downsized 4- reasons for investigating the reverse-uniflow: the interaction would stroke gasoline engine with a maximum engine speed of 6500 rpm. affect the exhaust phase in that iteration, not the intake as here. Modifying the profile to double its frequency whilst retaining the acceleration limit resulted in a minimum duration of 110° with 4 mm Varying the sleeve phase relative to the piston was also investigated of lift. Consequently, delaying EVO to increase ER resulted in EVC as part of this study; this changes the exhaust timing, leaving the occurring later in the compression stroke and the resulting loss of intake timing essentially piston-controlled. The optimum sleeve charge had then to be compensated by charging system work, which phase for the peak torque point (B) was found to be a 15° lead; was 3.4% higher than the piston-ported engine. Thus this concept however, for the part load (A) and maximum power (C) conditions a cannot match the late EVO of the OP2S design and has a higher 5° to 10° lag gave a slight improvement. One can imagine current ISFC. Also, from Table 6, the ER is less than the CR, meaning a camshaft phasing devices being able to provide this functionality, thermodynamic loss versus an ideal cycle. depending on the exact sleeve drive mechanism.

Page 15 of 21

7/20/2015 It must be stated again that for this and the reverse-uniflow sleeve- Nevertheless, after optimization this arrangement gave the highest valve case, heat transfer from the cylinder was assumed to be the average NSFC relative to all of the other concepts. The two principal same as for the other designs. There may be a small benefit to NSFC factors that limit this design with respect to the others are: from reduced heat transfer to the liner due to the presence of the sleeve, but at present this is unknown. For both sleeve-valve 1. Surface area. Although it has a smaller surface area due to the configurations this would be worthy of further work were either square stroke:bore ratio, the overall heat transfer is greater than concept to be carried forward, both in terms of theoretical analysis the other designs. Transfer to liner is less but to the piston and and, initially, motoring rig-based work. head it is greater due to the increased bore size.

Reverse-Uniflow Sleeve Valve 2. Port geometry. The exhaust and intake port dimensions are limited circumferentially since they lie on the same plane of the As mentioned above the investigation of this configuration was cylinder. To increase the area, the port height has to be driven by earlier findings rating the Crecy design versus the port- increased, resulting in earlier timing, reducing the expansion poppet configuration [34]. While the forward-uniflow sleeve-valve ratio and the work produced compared to the other concepts. optimization provided very similar results to the port-poppet, it was felt from constructing and operating the original models in that initial From this work it was found that the CTV in the exhaust port does work that it might be possible to provide increased expansion in a improve NSFC by allowing asymmetric timing, reducing the charge reverse-uniflow version of the sleeve valve layout. Furthermore, loss caused in conventional loop-scavenged engines by the unlike in the conventional Crecy arrangement, during scavenging any symmetrical and late closure of the exhaust; without the CTV this plug flow through the cylinder would also provide the advantage of occurs in all such engines even if they have variable port height leaving fresh air at the top (near any injector and position) mechanisms. This work therefore validates previous research and burnt gas at the bottom, helping to insulate the piston from performed by some of the authors [7,8,15,36]. The optimum point combustion heat loss. It is accepted, however, that the piston might was found to be with the CTV set to give EPC at 220° ATDC; be expected to run hotter in this case. however, it still underperformed the other concepts taken to this level of optimization at all operating points, although it is interesting to Table 5 shows that the reverse-uniflow sleeve has the second-best note that it was only slightly worse and it did not have the highest NSFC of the five concepts, after the OP2S. It does, however, have estimated supercharger power requirement. Considering its overall the highest estimated supercharger power requirement. The reason compactness and mechanical simplicity and adjustability, the concept that these two opposing results can co-exist is thought to lie in the may still be attractive for many applications. data shown in Table 6, where it can be seen that the original hypothesis is borne out: this configuration has an ER 23% greater Comparison of the different systems than its CR, meaning a significant thermodynamic advantage over the others in terms of work extraction (even the OP2S). It can thus be From Table 5 it can be seen that, at 188 g/kWh when averaged, the considered to be a bona fide Miller-cycle-operation engine, ironically NSFC of the OP2S is significantly better than the other concepts. without requiring all of the poppet-valve paraphernalia that permits Indeed, the NSFC results for the other four concepts are all quite existing 4-stroke engines to adopt this operating strategy. In fact, close, spanning 201-208 g/kwh (a range of 3.5%). Averaging these arguably it may better be termed a different form of Atkinson’s values together yields 205 g/kWh, meaning that the OP2S is better “Cycle Engine”, because it achieves its full operating cycle in a than the average of the rest by approximately 8.3%. single turn of its output shaft (although the original such concept actually operated on a 4-stroke cycle) [46]. It is believed that this This clear advantage for the OP2S stems from several areas, as possibility is a new finding for a pure 2-stroke cycle engine. mentioned above: reduced heat transfer, increased expansion work, and reduced supercharger power requirement, the latter two of which Since the intake port is at the top it is solely controlled by the sleeve are linked. They are an embodiment of the architectural advantage of motion and by adjusting the phase of the sleeve relative to the piston the OP2S in that it can use the whole cylinder bore circumference for the intake timing can be varied, allowing a range of values of the its ports (allowing for the lands between them), giving the optimizer over-expansion to be explored. Determining the exhaust timing is the opportunity to achieve the necessary angle-areas combined with more complicated due to the multiple interactions involved between short durations, yielding the related maximum expansion work. the sleeve, its ports and the piston. However, this design showed Because more work is generated in the cycle, this reduces the lower residual rates than the others and this in itself would be worth required air mass flow and with it the charging system work studying in more detail since it may allow yet more efficiency to be requirement, providing a virtuous circle in terms of the necessary realized from this concept as well as perhaps showing some synergies angle-areas on the intake and exhaust side. with turbocompounding. As mentioned above, the other four concepts are relatively closely Schnürle loop-scavenged piston-ported matched. However, the initial premise that the reverse-uniflow sleeve-valve configuration should be better than the others appears to As mentioned previously, the specific form of this engine that was be borne out, despite the fact that this configuration has the highest modelled was configured with the Lotus charge trapping valve in the supercharger power requirement. The unusual finding of this work is exhaust port because this was considered to represent the most that the reverse-uniflow sleeve-valve appears to be a 2-stroke advanced timing control device for a piston-ported engine. configuration which would allow differential expansion, and with it a Consequently, it would be expected to yield a best-case baseline for thermodynamic advantage (see Table 6). Of the others, only the the other configurations. OP2S is neutral in this respect, and again it is interesting to note that the remaining three all have very similar ratios of effective expansion

Page 16 of 21

7/20/2015 ratio divided by effective compression ratio, with values in the region of minimum NSFC in the design; however, in some cases the of 0.81-0.84. resultant timings showed significant differences to the values given by the guidelines, particularly for the blowdown phase. In turn this The overall results for the loop-scavenged engine shows the suggests that further research into angle-area requirements would be remarkable capacity of the CTV to improve scavenging performance; especially useful for the ongoing study of modern 2-stroke engines. from their publications Lotus always claimed that the CTV smoothed out the torque curve of the engines it was fitted to [7,8,15,36], and A subjective final ranking of the scavenging concepts, based on the this work provides insight into why this should be so. The fact that objective results of Table 5, is considered to be as shown in Table 7. this configuration has an average NSFC only 4 g/kWh worse than the commercially-successful (in terms of applications where fuel Table 7 Ranking of the different concepts based on the results obtained from consumption is considered of overriding importance) port-poppet this study. engine is considered remarkable, and a positive indictment of the CTV. It should be remembered that the simple non-CTV Schnürle Ranking Engine Type Comments loop-scavenged engine was not modelled, however; doing this would be worthwhile to provide some further context. 1st OP2S Clearly the best for NSFC and estimated required supercharger power. Also the least complicated mechanically. May be possible to apply VCR via Of the remaining port-poppet and forward-uniflow sleeve-valve crankshaft phasing. engines, the former is better for NSFC and worse for estimated supercharger power requirement. As a consequence of the 2nd Reverse- Allows differential expansion. May be better still with assumptions made in order to conduct this study it is tempting to rank uniflow further optimization. Has highest estimated required sleeve-valve supercharger power. Mechanically challenging due to them equally. current engineering knowledge.

In Table 5, the OP2S has overwhelmingly the lowest estimated 3rd = Port-poppet Compromised by valve kinematics. Obvious potential to supercharger power requirement. Logically, due to the more- vary exhaust valve timing. Not complex. beneficial ratio between turbine work and compressor work that would be expected for the whole scavenge air supply system, this 3rd = Forward- Mechanically challenging due to current engineering would also suggest that the OP2S might be the best suited to uniflow knowledge. sleeve-valve turbocompounding as well. While not investigated here, this has successfully been applied to 2-stroke engines in the past [21,44], with 3rd = Loop- Almost as good in NSFC as port-poppet and forward- perhaps the most famous example being the [47] scavenged uniflow sleeve, with better estimated required which, like Tresilian’s X-engine mentioned above, fell victim to piston-ported supercharger power than either. This performance was improvements in gas turbine engine performance. Nevertheless, it is greatly helped by the use of a Lotus-type charge trapping valve. Has validated potential for VCR. Mechanically interesting to note that the last reciprocating piston engines simple and understood. realistically to be considered as long-range aircraft powerplants were turbocompounded 2-stroke engines. The more recent work of Witzky 6 Poppet-valve Not taken to full optimization stage due to having the and co-workers is particularly pertinent in this respect as well [48]. reverse-loop poorest performance in the first stage of the process. The application of such technology should give further-improved fuel economy and is therefore considered worthy of further investigation particularly in connection with the opposed-piston and reverse- 5 uniflow sleeve valve engines . Other technologies which would be A ranking process for manufacturing was conducted as part of the interesting to study in this context might be the stepped-piston project, but this is not included here due to the two facts that this was concept, which could replace an external supercharger while still primarily an efficiency-oriented investigation and that such a ranking permitting turbocharging to be applied. This has been investigated is dependent on many more variables (for example, the sleeve valve extensively by Hooper [50,51] and also Lee [52] in a compounded manufacturing process being one complete unknown in the modern engine, as well as being a feature (in inverted form) of the engine of world). Nevertheless, it can still be said that the OP2S engine Witzky et al. [48]. continues to be very attractive when compared to the others in this manner, meaning that its overall advantage can be expected to Finally, from this work and also from an academic perspective, an increase further. important finding has been that the ‘standard’ guidelines for deciding the angle-area requirements of conventional crankcase scavenged 2- Final points with regards to this ranking are included in the stroke engines, as derived by researchers at Yamaha [43], have been Conclusions section below. found to be broadly applicable to all of the 2-stroke scavenging systems studied here. As discussed above, in the present work these guidelines were used as an initial starting point for port geometry, Conclusions however as the study progressed it also became clear that they were not fully optimal. Hence numerical optimization of the port/valve Several different scavenging systems suitable for use in an timings was adopted. This latter stage allowed the explicit targeting automotive 2-stroke engine were compared using a 1-D engine simulation code. The simulation was carried out on a single cylinder. All configurations were subject to the same indicated power and torque targets, themselves suitable for a medium-duty application. 5 More recently a large-capacity turbocompound V24 engine has , geometric CR and exhaust back pressure were been proposed to replace the gas turbines of passenger aircraft; all matched; however, engine stroke:bore ratios were chosen to allow however, the new proposal is for a 4-stroke engine [49]. a meaningful comparison, given the variation in porting arrangements. Page 17 of 21

7/20/2015 The conclusions drawn from this work were: Compared to the others, this is a challenge which expanded theoretical and rig work might help to mitigate. 1. The opposed-piston configuration provides the best performance since it allows for greater expansion than all of the 9. The poppet-valve reverse-loop was not taken to a full others (except the reverse-uniflow sleeve-valve) and the optimization; rather the concept was eliminated early on due to minimum heat transfer. This latter benefit occurred despite the its poor performance at that stage. This is not to imply that OP2S having the highest surface area-to-volume ratio of any of there are not other merits with this concept, such as the configurations tested; it was the result of the summations of commonality of production machinery with existing 4-stroke the product of surface area and HTC for the individual walls of engines, just that in terms of an approach prioritizing the fuel the combustion chamber. It gave an average NSFC 9.6% lower consumption potential of different scavenging systems it did than the loop-scavenged engine, which was found to be not perform as well as any of the others. marginally the worst performing layout. 10. Throughout this work, regardless of the configuration being 2. Varying the piston phasing of the OP2S engine was found to analyzed, it was found that existing experiential guidelines for have minimal effect on fuel consumption, at least within a port angle-area specification for loop-scavenged, piston-ported range of 2.5° to 12.5°. From previous work where VCR was engines using crankcase compression could also be applied to found to be a major control on an HCCI-type combustion them all. It is believed that this has not been demonstrated system in a 2-stroke engine [7,8], this could be a significant before, across such a broad variation of layout. However, even benefit if such a phasing mechanism can be engineered. This is using this approach, the numerical optimizer used also allowed considered worthy of further work. further improvements to be made. The paper therefore presents a fundamental comparison of scavenging systems using a new 3. The reverse-uniflow sleeve-valve engine was second in the approach, providing information which has not been shown final ranking of performance in terms of NSFC, at 3.4% lower before. than the loop-scavenged configuration. This was despite having the highest estimated supercharger drive power of all Recommendations for Further Work the systems in the complete study. Because of this performance in terms of NSFC it is considered worthy of further work This work has necessarily been constrained by the requirement to because this may reveal some potential for further-improved limit the number of variables in the analysis. It is recommended that fuel consumption. Its novelty itself makes this a possibility. for the most desirable configurations a sensitivity analysis be conducted to gauge the effect of different scavenging characteristics, 4. The poppet-valve uniflow approach was limited by the stroke:bore ratios, port timings, pumping work, heat transfer, and the kinematics of the valve train system. Changing to a system not combustion rates and phasings for the GCI combustion, to allow limited by valve springs would help in this area, but this was some assessment of the accuracy of the results. The investigation of beyond the scope of this project which adopted valve masses factors affecting heat transfer is very important should further and acceleration limits along with established current 4-stroke analysis of the sleeve-valve configurations be carried out, for practice. simplicity those having been assumed to be neutral with respect to the other arrangements here. Multi-cylinder engine configurations 5. The forward-uniflow sleeve-valve engine was considered to should be investigated too. have very good potential for breathing at higher engine speeds. This was slightly compromised by piston and sleeve motion Although the comparison reported here is based on net indicated interactions towards BDC, reinforcing the rationale to results, it would also be useful to show any sensitivity to the effect of investigate the reverse-flow version. frictional losses. This is particularly the case for the OP2S engine where the crankshaft timing mechanism will potentially introduce a 6. In order to investigate both the forward- and reverse-uniflow significant penalty and where the frictional losses associated with the configurations with the sleeve-valve, layout drawings and two crankshafts will be different as a result of the exhaust piston analysis of the Crecy-type sleeve had to be undertaken, using producing more power than the intake [53]. design details gleaned from the few remaining documents pertaining to this engine (essentially, the Rolls-Royce Heritage Trust book on the subject [21]).

7. The loop-scavenged engine performed essentially as well as the port-poppet and forward-uniflow sleeve-valve engines. However, this is principally due to the adoption of the Lotus charge trapping valve device, modified for use here from Lotus’s publications. As a result of its mechanical simplicity, compactness, and the familiarity of its basic arrangement, it may arguably be considered more attractive than all of the others except the OP2S.

8. Heat transfer was not modified for either of the sleeve-valve engines, which, from a modern perspective, would demand most design and developmental time being spent on them.

Page 18 of 21

7/20/2015 Contact Information 7. Turner, J.W.G., Blundell, D.W., Pearson, R.J., Patel, R., Larkman, D.B., Burke, P., Richardson, S., Green, N.M., Brewster, S., Kenny, R.G., and Kee, R.J., “Project Omnivore: A Dr James Turner Variable Compression Ratio ATAC 2-Stroke Engine for Ultra- Professor of Engines and Energy Systems Wide-Range HCCI Operation on a Variety of Fuels”, SAE Department of Mechanical Engineering technical paper 2010-01-1249 and SAE Int. J. Engines pp. 938- University of Bath 955, 2010. Claverton Down 8. Blundell, D.W., Turner, J.W.G., Pearson, R.J., Patel, R., and Bath BA2 7AY Young, J.J., “The Omnivore Wide-range Auto-Ignition Engine: United Kingdom Results to Date using 98RON Unleaded Gasoline and

Fuels”, SAE technical paper 2010-01-1249 and SAE Int. J. [email protected] Engines pp. 938-955, 2010.

9. Regner, G., Herold, R.E., Wahl, M.H., Dion, E., Redon, F., Robert Head Johnson, D., Callahan, B.J. and McIntyre, S., “The Achates Senior Laboratory Scientist Power Opposed-Piston Two-Stroke Engine: Performance and Fuel technology Division Emissions Results in a Medium-Duty Application”, SAE Research and Development Center technical paper 2011-01-2221 and SAE Int. J. Engines, Vol. 4 , Saudi Arabian Oil Company Iss. 3, pp. 2726-2735, 2011, doi:10.4271/2011-01-2221. PO Box 13563 10. Redon, F., Kalebjian, C., Kessler, J., Rakovec, N., Headley, J., Dhahran 31311 Regner, G., and Koszewnik, J., “Meeting Stringent 2025 Saudi Arabia Emissions and Fuel Efficiency Regulations with an Opposed-

Piston, Light-Duty Diesel Engine”, SAE technical paper 2014- [email protected] 01-1187, SAE 2014 World Congress, Detroit, Michigan, USA, 8th–10th April, 2014, doi:10.4271/2014-01-1187. Acknowledgments 11. Sharma, A. and Redon, F., “Multi-Cylinder Opposed-Piston Engine Results on Transient Test Cycle”, SAE technical paper Funding from Saudi Aramco to conduct this research is very 2016-01-1019, SAE 2016 World Congress, 12th-14th April, gratefully acknowledged, as is the permission given by the Rolls- 2016, doi:10.4271/2016-01-1019. Royce Heritage Trust to use illustrations from two of their 12. Naik, S., Johnson, D., Fromm, L., Koszewnik, J., Redon, F., publications pertaining to the Crecy engine. Regner, G., and Abani, N., “Achieving Bharat Stage VI Emissions Regulations While Improving Fuel Economy with the Discussions with Professor Chris Brace of the University of Bath Opposed-Piston Engine”, SAE technical paper 2017-26-0056 were also fundamental in the choice of the approach adopted and in and SAE Int. J. Engines 10(1):2017, doi:10.4271/2017-26-0056. the decisions that were made during the project. 13. Ishibashi, Y. and Asai, M., “Improving the Exhaust Emissions of Two-Stroke Engines by Applying the Activated Radical Combustion”, SAE technical paper 960742, SAE 1996 World Finally, useful discussions with David Robinson of Blacklock & Congress, Detroit, Michigan, USA, 26th-29th February, 1996, Bowers were conducted regarding aero engine applications and doi: 10.4271/960742. turbocompounding in general. 14. Benajes, J., Novella, R., De Lima, D., and Tribotté, P., “Investigation on Multiple Injection Strategies for Gasoline PPC References Operation in a Newly Designed 2-Stroke HSDI Compression Ignition Engine”, SAE technical paper 2015-01-0830 and SAE 1. Sher, E., “Scavenging the 2-stroke engine”, Prog. Energy. Int. J. Engines 8(2):2015, 2015, doi:10.4271/2015-01-0830. Combust. Sci. Vol. 16, pp. 95-124, 1990. 15. Blundell, D., Turner, J.W.G., Duret, P., Lavy J., Oscarsson, J., 2. Torrens, H., “Joseph Day 1855-1946 and the Development of Emanuelsson G., Bengtsson, J., Hammarström, T., Perotti, M., the Two-Stroke Internal-Combustion Engine”, Bath Industrial Kenny, R., and Cunningham, G., “Design and Evaluation of the Heritage Trust Limited, Bath, UK, 1991, ISBN 0 9518602 0 8. ELEVATE Automotive Two-Stroke Engine”, SAE technical 3. Schlunke, K., “The Orbital Combustion Process Engine”, 10th paper 2003-01-0403, SAE 2003 World Congress, Detroit, Vienna Motor Symposium, Vienna, Austria, 27th-28th April, Michigan, USA, 3rd-6th March, 2003, doi:10.4271/2003-01- 1989. 0403. 4. Shawcross, D., Pumphrey, C., and Arnall, D., “A Five-Million 16. https://en.wikipedia.org/wiki/Knight_engine, last accessed 26th Kilometre, 100-Vehicle Fleet Trial, of an Air-Assist Direct Fuel July, 2018. Injected, Automotive 2-Stroke Engine”, SAE technical paper 17. Turner, J.W.G. and Lewis Monsma, J., “An investigation into 2000-01-0898, SAE 2000 World Congress, 6th-9th March, the port timing of a Burt-McCollum sleeve valve and its 2000. interaction with a simple variable compression ratio 5. Hundleby, G.E., “Development of a Poppet-Valved Two-Stroke mechanism”, SAE technical paper 2017-24-0168, 13th Engine - The Flagship Concept”, SAE technical paper 900802, International Conference on Engines and Vehicles, Capri, Italy, SAE 1990 World Congress, Detroit, Michigan, USA, 26th 10th-14th September, 2017. February to 2nd March, 1990. 18. Hassell, P., “The Bristol Sleeve Valve Aero Engines”, Chapter 6 6. Stokes, J., Hundleby, G.E., Lake, T.H., and Christie, M.J., of 'The Piston Engine Revolution' Conference of the Newcomen “Development Experience of a Poppet-Valved Two-Stroke Society, pp. 112-132, Manchester, UK, 14th-17th April, 2011. Flagship Engine”, SAE technical paper 920778, SAE 1992 19. Ricardo, H.R., The High-Speed Internal-Combustion Engine, 3rd World Congress, Detroit, Michigan, USA, 24th-28th February, Edition, Blackie & Son Ltd., , UK, 1941. 1992.

Page 19 of 21

7/20/2015 20. Jones, L., Sectioned drawings of Piston Aero Engines, Historical Detroit, Michigan, USA, 12th-14t April, 2016, Series Special Edition, Rolls-Royce Heritage Trust, Derby, UK, doi:10.4271/2016-01-0659. 1995, ISBN 1 872922 07 4. 39. Warey, A., Andruskiewicz, P., Durrett, R., Gopalakrishnan, V., 21. Nahum, A., Foster-Pegg, R.W., and Birch, D., The Rolls-Royce Potter, M., and Najt, P., “CO2 Emissions Benefits of Alternative Crecy, Rolls-Royce Heritage Trust Historical Series No. 21, Diesel Engine Architectures”, 27th Aachen Colloquium, Book 2, Rolls-Royce Heritage Trust, Derby, UK, 1994, ISBN 1 872922 pp. 901-912Aachen, Germany, 9th-10th October, 2018. 05 8. 40. Mattarelli, E., Rinaldini, C., Savioli, T., Cantore, G. , Warey, A., 22. Smith, P.H., edited Setright, L.J.K, Valve mechanisms for high- Potter, M., Gopalakrishnan, V., and Balestrino, S., "Scavenge speed engines - Their design and development, 2nd ed., G.T. Ports Optimization of a 2-Stroke Opposed Piston Diesel 23. Setright, L.J.K., Some Unusual Engines, Mechanical Engine," SAE technical paper 2017-24-0167, SAE 2017 World Engineering Publishing Ltd, London, UK, 1975, ISBN 0 85298 Congress, Detroit, Michigan, USA, 4th-6th April, 2017, 208 9 doi:10.4271/2017-24-0167. 24. White, G., Allied Aircraft Piston Engines of World War II, 41. Laget, O., Ternel, C., Thiriot, J., Charmasson, S., Tribotté, P., Society of Automotive Engineers, Inc., Warrendale, and Vidal, F., "Preliminary Design of a Two-Stroke Uniflow Pennsylvania, USA, 1995, ISBN 1-56091-655-9. Diesel Engine for Passenger Car," SAE technical paper 2013- 25. Bingham, V.F., Major Piston Aero Engines of WWII, The 01-1719 and SAE Int. J. Engines 6(1):2013, doi:10.4271/2013- Crowood Press Ltd, Marlborough, UK, 1998, ISBN 978- 01-1719. 1840370126. 42. Tribotté, P., Ravet, F., Dugue, V., Obernesser, P., Quechon, N., 26. https://en.wikipedia.org/wiki/Junkers_Jumo_205, last accessed Benajes, J., Novella, R., and De Lima, D., “Two Strokes Diesel 26th July, 2018. Engine - Promising Solution to Reduce CO2 Emissions”, 27. Pearce, W., “Junkers Jumo 223 ”, Old Machine Procedia - Social and Behavioral Sciences Vol. 48, pp. 2295- Press, available at 2314, 2012, doi: 10.1016/j.sbspro.2012.06.1202. https://oldmachinepress.com/2015/09/26/junkers-jumo-223- 43. Naitoh, H. and Nomura, K., "Some New Development Aspects aircraft-engine/, last accessed 26th July, 2018. of 2-Stroke Cycle Motorcycle Engines," SAE technical paper 28. Pearce, W., “Junkers Jumo 224 Aircraft Engine”, Old Machine 710084, SAE 1971 World Congress, Detroit, Michigan, USA, Press, available at February, 1971, doi:10.4271/710084. https://oldmachinepress.com/2015/10/03/junkers-jumo-224- 44. Chatterton, E., “The Napier Deltic diesel engine”, SAE technical aircraft-engine/, last accessed 26th July, 2018. paper 560038 and SAE 1956 Transactions, Vol. 64, pp. 408-425, 29. https://en.wikipedia.org/wiki/Detroit_Diesel_Series_71, last 1956. accessed 26th July, 2018. 45. Pirault, J.-P. and Flint, M., “Opposed Piston Engines - 30. https://marine.mandieselturbo.com/applications/projectguides/2s Evolution, Use, and Future Applications”, SAE International, troke/content/printed/46-26mc.pdf, last accessed 26th July, 2018. Warrendale, PA, USA, SAE Order No. R-378, ISBN 978-0- 31. https://marine.mandieselturbo.com/two-stroke/2-stroke- 7680-1800-4, 2010. engines/me-b-engines, last accessed 26th July, 2018. 46. https://en.wikipedia.org/wiki/Atkinson_cycle#Atkinson_%22Cy 32. https://en.wikipedia.org/wiki/W%C3%A4rtsil%C3%A4- cle_Engine%22, last accessed 8th August, 2018. Sulzer_RTA96-C, last accessed 26th July, 2018. 47. Sammons, H. and Chatterton, E., “The Napier Nomad Aircraft 33. Wang, X., Ma, J. and Zhao, H., “Evaluations of Scavenge Port Diesel Engine”, SAE Transactions Vol. 63, No. 107, pp. 107- Designs for a Boosted Uniflow Scavenged Direct Injection 131, 1955. Gasoline (BUSDIG) Engine by 3D CFD Simulations”, SAE 48. Witzky, J.E., Meriwether, R.F., and Lux, F.B., “Piston-Turbine- technical paper 2016-01-1049, SAE World Congress, Detroit, - A Design and Performance Analysis”, SAE Michigan, USA, 12th-14th April, 2016, doi: 10.4271/2016-01- technical paper 650632, SAE International West Coast Meeting, 1049. Vancouver, B.C., Canada, 16th-19th August, 1965. 34. Turner, J.W.G., Head, R.A., Wijetunge, R., Chang, J., Engineer, 49. Eilts, P. and Friedrichs, J., “Investigation of a Diesel Engine for N., Blundell, D., and Burke, P., “Analysis of different uniflow Aircraft Application”, AIAA Propulsion and Energy Forum, scavenging options for a medium-duty 2-stroke engine for a Atlanta, Georgia, USA, 10th-12th July, 2017. U.S. light-truck application”, technical paper ICEF2018-9766, 50. Hooper, P.R., “Investigation into a stepped piston engine ASME 2018 Internal Combustion Engine Fall Technical solution for automotive range- extender and hybrid electric Conference, San Diego, California, USA, 4th-7th November, vehicles to meet future green transportation objectives”, 2018. Proceedings of the Institution of Mechanical Engineers Part D: 35. Blair, G.P., Design and Simulation of Two-Stroke Engines, SAE Journal of Automobile Engineering Vol. 232(3), pp. 305-317, International, Warrendale, PA, USA, 1996, ISBN 1560916850. 2018, doi: 10.1177/0954407017698304. 36. Blundell, D.W. and Sandford, M.H., “Two-Stroke Engines - The 51. Hooper, P.R., Novel Three-Cylinder Engine Solutions Offering Lotus Approach”, SAE technical paper 920779, SAE 1992 Low Noise Vibration and Harshness for Range-Extender and World Congress, Detroit, Michigan, USA, 24th-28th February, Hybrid Electric Vehicles”, SAE technical paper 2018-01-1553, 1992, doi:10.4271/920779. 10th International Styrian Noise, Vibration & Harshness 37. Badra, J.A., Sim, J., Elwardany, A., Jaasim, M., Viollet , Y., Congress: The European Automotive Noise Conference, Graz, Chang, J., Amer, A., and Im, H.G., “Numerical Simulations of Austria, 20th-22nd June, 2018, doi:10.4271/2018-01-1553. Hollow-Cone Injection and Gasoline Compression Ignition 52. Lee, C.-P., “Turbine-compound free-piston linear Combustion With Naphtha Fuels”, Journal of Energy Resources engine”, Ph.D. thesis, University of Michigan, Ann Arbor, Technology, Vol. 138, Issue 5, February 2016, doi: Michigan, USA, 2014. 10.1115/1.4032622. 53. Morton, R., Riviere, R., and Geyer, S., “Understanding Limits to 38. Warey, A., Gopalakrishnan, V., Potter, M., Mattarelli, E., and the Mechanical Efficiency of Opposed Piston Engines”, SAE Rinaldini, C.A., “An Analytical Assessment of the CO2 technical paper 2017-01-1026, SAE 2017 World Congress, Emissions Benefit of Two-Stroke Diesel Engines”, SAE Detroit, Michigan, USA, 4th-6th April, 2017, doi:10.4271/2017- technical paper 2016-01-0659, SAE 2016 World Congress, 01-1026. Page 20 of 21

7/20/2015 FC Fuel cell

GCI Gasoline compression ignition Definitions/Abbreviations HCCI Homogeneous charge 1-D One-dimensional compression ignition

AFR Air-fuel ratio HTC Heat transfer coefficient

ATDC After top dead centre ICE Internal combustion engine

BDC Bottom dead centre IMEP Indicated mean effective pressure BMEP Brake mean effective pressure IPC Inlet port closing (timing)

BTDC Before top dead centre IPO Inlet port opening (timing)

CI Compression ignition OEM Original equipment manufacturer CR Compression ratio PEM Proton exchange membrane CTV Charge trapping valve Px Exhaust manifold pressure DI Direct injection SI Spark ignition EPC Exhaust port closing (timing) TDC Top dead centre EPO Exhaust port opening (timing) VCR Variable compression ratio

EV Electric vehicle VVT

EVO Exhaust valve opening Mechanical Mechanical efficiency (timing)

ER Expansion ratio

Page 21 of 21

7/20/2015